00:02.06 Ratnadeep Bhattacharjee Welcome to another episode of Leaders Perspective, where we sit down with healthcare care and technology leaders driving real change in the industry. I'm your host, Ratnadeep Bhattacharji, co-founder at TechVariable. 00:13.84 Ratnadeep Bhattacharjee Today's topic is ah why healthcare care AI needs interoperability to succeed. You know, as AI continues to transform healthcare, care one truth remains, AI can only be as powerful as the data it can access, understand, and probably trust. 00:30.89 Ratnadeep Bhattacharjee Interoperability is no longer a nice to have. It's a critical enabler of innovation and meaningful outcomes. Joining me today is John Lean, one of the most recognized voices in the healthcare IT media. 00:43.82 Ratnadeep Bhattacharjee John is the founder and chief editor at Healthcare Scene, which includes the Healthcare IT Today and Swy.Health communities, publications, podcasts and conferences, you name it. 00:54.82 Ratnadeep Bhattacharjee Healthcare IT Today is the leading healthcare IT community with over 19,000 published articles, 1,000 plus videos and audio podcasts and a community of 200,000 followers. 01:07.68 Ratnadeep Bhattacharjee Swidoth Health is a unique platform where healthcare B2B marketers in hospitals and health systems marketers you know collaborate to improve patient lives. Beyond his media work, John also advises healthcare IT companies and has built a strong personal presence as a tech guy on Twitter or X. John, it's a pleasure to have you here. 01:29.07 Ratnadeep Bhattacharjee Welcome to Lita's Perspective. 01:30.95 John Lynn Yeah, thanks so much for having me. I'm excited for today's discussion. 01:36.22 Ratnadeep Bhattacharjee Great. Great. John, ah to start off, you know you have been covering healthcare IT for years and um have seen probably trends come and go than most even healthcare executives, honestly. 01:48.14 Ratnadeep Bhattacharjee How would you describe and the current relationship between know all the AI hype, AI innovation and the most important part of all of it, you know inter interoperability maturity in healthcare today? How does the relationship between both of these, ah according to you, 02:07.02 John Lynn Yeah, well, I think it is the hypiest hype that you could ever have. ah But I also think it's based on actual ah things being solved and solutions to problems we never thought we could solve. 02:23.51 John Lynn So that's the exciting part. I often have said that it's the most exciting thing since meaningful use, which was $36 million of stimulus money from the government. right But it turns out we are don't even have money from the government and people are just as excited. So you know I think that you know from a hype cycle, it's definitely one of the the biggest trends and changes in industry. I you know, a former CIO that I was talking to at a conference recently, she said, John, how do you feel about the conference? And I asked her the same question. 02:56.65 John Lynn And she said, you know what, for me, every single company is leveraging AI to redo what they're do, what what they the problem they solve or to rethink how they solve the problem. 03:08.70 John Lynn She's like, before you would go to a vendor and you're like, oh yeah, that vendor does faxing or that vendor does you know call center agents or that vendor does EHR whatever it might be. 03:19.62 John Lynn She's like, and every one of them is now innovating because of what's possible because of AI. So I think that's kind of a groundwork for the hype. 03:30.16 John Lynn As far as the relationship between interoperability, you highlighted it perfectly. If you do not have good data, your AI will not be good. like That's it. like 03:40.95 Ratnadeep Bhattacharjee Thank you. 03:41.44 John Lynn If you put the you it's bad data in, it means to bad outcomes out. right i mean This is just simple 101, but it's the reality. and it's What's interesting is it's the reality that healthcare care organizations face because as they implement the AI solutions, they're finding out that the emperor may not have any clothes on. to right like that they That their data is not as pretty as they think it is. And so then they're going back and saying, i got to fix my data. I got to add more data to make it more valuable because I never did it before. So I think that's kind of where we're at with the trends is like, hey, we need good data and not just data, but trusted data that we can apply these AI models to, to ensure that we get the outcomes that we want. 04:34.49 Ratnadeep Bhattacharjee and Interesting. ah John, you you you pointed out the fact that, you know, in AI or whatever you do in terms of innovation, right? Or, you know, for that matter, whatever you do with your data, right? 04:47.64 Ratnadeep Bhattacharjee Garbage in is garbage out, rightly pointed out. So AI thrives on quality data. But in healthcare, ah much of the data is locked in silos. forty um and um so Most of these are unstructured data or a combination of structured and structured data. 05:04.68 Ratnadeep Bhattacharjee Sometimes data is incomplete as well. So from your vantage point, you know speaking with multiple healthcare leaders and having seen the healthcare landscape for for for more than and know two decades now, how big of a barrier is data fragmentation to AI success? 05:22.62 Ratnadeep Bhattacharjee and what needs to change to truly, you know, kind of unlock its potential. 05:28.74 John Lynn You know, it's interesting because there are slices of AI that don't even need the data. and You know, AI medical scribes, they've done incredible work without the data. It's going to be interesting as they get more data, how they're going to improve even more. 05:43.87 John Lynn But they've solved a substantial problem, the documentation burden, and they really haven't needed the data because they're getting the raw data from from the visit by recording it, right? So there are workarounds that you can do that actually solve the problem in a really powerful way. 06:03.33 John Lynn But for a lot of the things, especially when you're looking at value-based care, you know taking care of the entire population population, understanding what's influencing the outcomes that are existing in your healthcare organization, when you look at that, you need more data. 06:19.94 John Lynn Now, you and you're 100% right. There are silos of data throughout healthcare. But I always like to remind people, like, silos are silos for a reason. 06:31.77 John Lynn And it's usually not because they can't get it out. ah It's usually because there's some sort of competitive pressure or there's no benefit that they've seen to opening up that silo. 06:37.50 Ratnadeep Bhattacharjee Mm. 06:46.67 John Lynn or Or maybe they just have a mental block that says, i need to do this to stay competitive with this. you know, my customer, my my competitor across the street. Right. And, and so the silos exist because many people don't want to share data or they're afraid to share data. And so I think, you know, that's one thing that's really important to understand is that, you know, if we wanted to solve the problem of data sharing and healthcare, we could, 07:17.99 John Lynn But there are a lot of groups, a lot of stakeholders in the healthcare industry that don't really want to share data. And so I think that's where it's a challenge. 07:29.15 John Lynn You know, the the other challenge with it is that healthcare data is quite complex. Like, we're not just talking about financial data. I heard someone say that healthcare data is an order of magnitude more complex than financial data. 07:42.32 John Lynn And when I heard him say that, I said, i think it's even worse than that. You know, like, financial data is quite tame in comparison to healthcare data and how it's done. 07:46.21 Ratnadeep Bhattacharjee Absolutely. 07:54.23 John Lynn Plus, there are perverse incentives in healthcare care that make people not want to share, even though that may be what's best for the patient. And so, you know, understanding the business of healthcare is it going to be key to unlocking the data and being able to share the data where it needs to go. 08:13.04 John Lynn Because what we've found is when there is a will, when there is a business reason to share data, then there is a way to do it and it will happen. 08:21.35 Ratnadeep Bhattacharjee and 08:24.74 Ratnadeep Bhattacharjee Okay, okay. but this is yeah This is unique, John, know, most of the people I talk to kind of give me generic responses, but this is really looking at both the sides of the coin. It's not a simple, it's not always a simple answer, It's not a yes or a no and in every aspect of, especially in healthcare, it's it's never that easy. yeah You know, like as someone who has worked in this space for the past seven to eight years, I've seen everything, right? 08:50.91 Ratnadeep Bhattacharjee So, I totally agree with you. 08:53.89 John Lynn Well, it's kind of sad too, right? 08:54.22 Ratnadeep Bhattacharjee john 08:55.26 John Lynn Like you think it should be like, okay, whatever's best for the patient, let's do that. 08:56.06 Ratnadeep Bhattacharjee yeah 09:01.72 John Lynn And sadly it's not, right? Like it's just the it's just not because healthcare care is a business, yeah you know, not to sidebar too much, but like when they say that healthcare is the number one employer in every, you know, 40 of the 50 states in the U.S. or so, it's something like that, right? I'm like, that's a really bad thing. 09:22.65 John Lynn Because that means so many people's livelihoods depend on health care. 09:22.94 Ratnadeep Bhattacharjee Hmm. 09:28.35 John Lynn And so if you say, hey we need reduce the cost of health care, guess what all those people employed in health care think? Wait, are you coming after my job? 09:35.33 Ratnadeep Bhattacharjee Absolutely. 09:36.29 John Lynn And they're like, no, we don't want to do that. Right. And so I think that that's where it it has so many of those perverse incentives. 09:39.90 Ratnadeep Bhattacharjee Hmm. Interesting. ah But John, you know, have you kind of seen examples like, for example, you know, longitudinal patient record, you know, people keep talking about it, you know, but as I can see, it's not that easy, you know, longitudinal patient record is still, ah ah you know, let's put it this way, it's still a dream for many people that, you know, 10:11.03 Ratnadeep Bhattacharjee accessing patient record or having a patient record just like you want it to be is not practically i would say it's possible but it's not yet achievable let's put it that way right so have you in in in all your discussions with different executives healthcare care executives and you know having seen most of the systems, healthcare systems, right? 10:33.61 Ratnadeep Bhattacharjee Have you seen examples where, you know, AI solutions in healthcare succeeded or even failed, right? Because of, especially because of interoperability issues, you know, what lessons can innovators take from these experiences? 10:50.21 John Lynn You know, your comment's interesting. i actually, over all these years, pretty strongly believe that having a longitudinal complete patient record is a pipe dream that will never happen. 11:07.68 John Lynn Like I just think when you look at where data is stored in healthcare, it's just going to be nearly impossible to reach the point where you can have all of the data for the patient in a full longitudinal record. 11:07.82 Ratnadeep Bhattacharjee wow 11:23.75 John Lynn Like it's it's a nice idea. and I kid you not, like every every month, every couple of weeks, I hear someone that comes and pitches me this amazing startup idea they had. 11:35.53 John Lynn They said, imagine how wonderful it would be if we had all the patient data in one record and the patient could give it to the doctor, the doctor could get access to it, whatever, right? And they're like, isn't this an amazing idea? 11:47.84 John Lynn and I just shake my head because I've been pitched it hundreds of times, right? like and So everyone agrees that if we could do that, it would be good, right? like There's no disagreements there. 11:58.61 John Lynn What they don't understand is that the the magnitude of the problem is so huge that I just don't think it will happen. 12:09.07 John Lynn And there's so many different players in the space. I mean, look at 12:12.64 Ratnadeep Bhattacharjee and 12:14.00 John Lynn your wearable vendors, look at your pharmacy vendors, look at your lab vendors, let alone the doctor and the practice and the surgery center and the urgent care and the telehealth and the whatever else, right? Like you have to bring all of those together with all of the data in some format. 12:31.08 John Lynn Let's not even talk about standards, right? that Maybe AI can solve this, the standards problem, but it's like, you know, you have to get all of those people together to be able to do it. 12:36.20 Ratnadeep Bhattacharjee it 12:40.11 John Lynn So To me, that's why ah i'm not sure it'll be solved. And then you layer on the other thing, which is, okay, if we do solve that problem, what... extra benefit are we going to get beyond what we already experienced today without a full longitudinal record? 12:58.65 John Lynn And that's, this that you know, as for a startup company, that's the problem they have to solve. 12:59.14 Ratnadeep Bhattacharjee Hmm. 13:03.42 John Lynn And when you look at it, you're like, well, it's not actually that helpful to a patient to be like, if I were to aggregate all my data, how would my life change? 13:14.64 John Lynn ah it wouldn't change right now. 13:15.15 Ratnadeep Bhattacharjee Hmm. 13:16.86 John Lynn If I'm a very chronic patient, it would. And that's true. But how that's a small part of the population. Anyway. So to me, I don't know if that is, but here's the other bigger idea. 13:28.91 John Lynn Do we even need that? Like, why do we need the full longitudinal patient record? Sure, it would be great if we could. and if we could, everyone would leverage it and you know in their solutions. But the nice thing is that we do have pockets of interoperability that are making a big difference in what's happening. 13:46.79 Ratnadeep Bhattacharjee Hmm. 13:47.41 John Lynn I mean, just look at prescriptions. Prescriptions is the most interoperable piece of healthcare right now. Some might argue it's because of a monopoly called SureScripts, right? That essentially, you know, shares all of these prescriptions and we'll see how that evolves. There's a few competitors. There were some... 14:05.84 John Lynn some legal cases and stuff that are opening now. But like it you know we have that data. You know who what prescriptions most people have taken. There's lots of nuance. There's lots of things that they could do to improve it. 14:17.61 John Lynn But like that's a place in interoperability that's doing really well. And we see that in other places. I mean, the reality is billions of transactions and sharing of health healthcare care data happens every month. 14:30.81 John Lynn across networks, across HIEs, across you know networks like SureScripts, HL7 messages. think it's billions and billions of HL7 messages. So interoperability is happening. 14:44.55 John Lynn The problem is this. We have billions of transactions, but what if the scale of the problem is it should be multi-trillions of transactions? Well, then we're barely have touched the problem, right? 14:55.27 Ratnadeep Bhattacharjee um 14:57.68 John Lynn So, you know, this is the problem with sharing data is that we could do it in in ah in a massive way. ah You know, getting more back to what you were saying, though, as far as like interoperability and shaping AI solutions, ah you know, i think the first one I already kind of highlighted where the AI says, i can't trust your data, so I'm going to collect it myself. 15:21.37 John Lynn And we're seeing that in the exam room with ambient you know voice technologies that are essentially capturing the entire exam room messaging. We're seeing it with things on the front end. 15:32.99 John Lynn So they're doing the intake using AI to ask the questions that are needed to then inform the visit in a different way. And the nice thing is that could actually be personalized. 15:44.78 John Lynn to an individual patient, it's not just a form sent to them, but then the patient fills out. Like the AI could actually say, wait, you told me there's a, you're you know, you have suicidal ideation. 15:56.35 John Lynn Maybe I could do something with that, right? course, I probably should make you have your call someone. But, you know, like, but you get the idea, right? If I find and discover that, you know, you're close to, you know, diabetes, 16:09.30 John Lynn Could I have that front end AI chat bot that's doing the intake ask a lot more questions than I would of someone who doesn't you know present with those symptoms? So you know that's another example of them collecting it on their own. 16:24.72 John Lynn The other part we see is a lot more organizations are layering it on top of the EHR. Now, we need the EHR to free up the data a bit more than they are, but we see ah some really interesting solutions. 16:31.09 Ratnadeep Bhattacharjee um 16:37.86 John Lynn the Maybe the most interesting right now that's really being used is the the AI chart summaries. So they're taking this plethora of data that's in the EHR across multiple lab results and multiple visits and all of those things and saying, can you summarize it for me? 16:48.15 Ratnadeep Bhattacharjee um 16:58.34 John Lynn And we're seeing this happen across a number of different ways. It will be interesting to see who wins out in this regard. But that's another example of them taking the data. They even might take HIE data. 17:09.75 John Lynn They may pull in that data you know from another source and then say, let's summarize it because the doctor doesn't have time to read through the thousand pages of of clinical data. 17:10.31 Ratnadeep Bhattacharjee Yeah. 17:19.60 John Lynn So thats that's another example that has been a huge success. right They've been able to get the data and they can do really cool things with it. you know One other that I'll just throw out there, it's amazing what they're doing with call center agents. 17:32.83 John Lynn like the AI call center agents. And we're seeing two forms. One is the fully automated one that's based on data from the practice so that they can answer the questions. 17:38.42 Ratnadeep Bhattacharjee yeah 17:41.94 John Lynn Like what insurance are you taking or what, you know all those questions that you might have or even scheduling the appointment, right? They're they're doing that kind of thing. But then we're also seeing the AI call center agent augment the call center agent by collecting all of the data. I mean, I've seen one from ECW, you know, they have this a Hilo Genie is what it's called. 18:04.78 John Lynn And it sits there and listens listens to the call. And then it actually prompts you for information from the EHR that could help you answer the question of the patient. It's like that is a beautiful use of AI with the right data can create a a magical experience that actually augments the call center agent. 18:27.15 Ratnadeep Bhattacharjee Interesting. and I mean, I, there are, there are plethora of use cases, John. I mean, we are just scratching the surface. I think the amount of work that has been done on the AI front is, is unprecedented, you know, so with the advent of AI or, I mean, AI took shape a number of years ago, right? 18:35.26 John Lynn Exactly. 18:49.11 Ratnadeep Bhattacharjee But with the advent of Gen AI and you know, agent EKI recently, you know, things have really taken an upturn, you know, innovation is not the problem. The problem is more from the perspective of enterprise outcomes and whether those outcomes really make sense, right? 19:03.90 John Lynn Thank you. 19:06.58 Ratnadeep Bhattacharjee Some outcomes really don't make sense. and And so have you, have you ah kind of encountered cases, John, wherein, you know, people talk a lot about, you know, doing this in agent EKI? 19:18.78 Ratnadeep Bhattacharjee Yeah, doing this, which is NTKI, doing that with the use of, you know, large language models, small language models, whatever you call it. And they are not really making sense when ah addressing the core issue here. 19:33.88 Ratnadeep Bhattacharjee They are just talking about the how how cool my model is, let's say, how cool my workflow is. But it's not really solving anything in the long run. Have you encountered anything like that, John? 19:47.40 John Lynn I mean, I'm sure there are some of those. i you know i think we've learned from the past you know that it's like, I'm not just implementing cool technology. i better be a big good business partner. 19:59.41 John Lynn Otherwise, i'm going to lose my job. 19:59.69 Ratnadeep Bhattacharjee Hmm. Hmm. 20:01.41 John Lynn you know i So I think we've seen ah and the credit that these leaders, right? That we have seen this evolution beyond just hey, I think this is cool, I want to implement it So I actually don't think we see that as much anymore. 20:17.37 John Lynn but you know What's interesting though is one you are 100% right. The biggest challenge if I'm a CIO in healthcare care today is there's too many opportunities. 20:30.47 John Lynn And so I have to look at it and say, i can't do everything 20:30.62 Ratnadeep Bhattacharjee Mmm. 20:35.20 John Lynn But what should I do now that's going to align with our strategic priorities? So I think that's the biggest challenges that CIOs face. And then the second one is this is so new. 20:47.58 John Lynn Often the data isn't there to tell you exactly what the outcome is going to be. Plus, it's moving so quickly. So even if I talk to my peer, hey, you implemented this, what did you think and how did it how did it go? 21:02.41 John Lynn Well, six months later, the solution has evolved so quickly like that that six month data is not even almost relevant to what's happening. Now, hopefully they've kept up and they're implementing it quickly and you can still get some data, but you get the idea, right? 21:18.32 John Lynn And so, ah you know, what I hear most CIOs saying is when I go into implementing this AI solution, 21:19.05 Ratnadeep Bhattacharjee right 21:26.15 John Lynn I have to have a clear vision of what the value prop is that I want. And then I have to incorporate the measurements to know if I'm actually achieving the value that I intended when i we first started. 21:41.06 John Lynn And I need to measure that rigorously And then I need to be held accountable to some AI committee or a tech governance committee to say, let me report on the outcomes. 21:51.35 John Lynn Because if it didn't, then let's scrap it, right? like and let's or let's figure out some other solution or or see what it is. The other problem is this, and I saw this at a class Arch Collaborative event. 22:03.81 John Lynn They call it, ah you know, being stuck in pilot purgatory, which is we do all these pilots and then we never get out of the pilots and we never scale it to the whole whole whole organization. 22:08.45 Ratnadeep Bhattacharjee Um. 22:15.02 John Lynn Anyway, go check out. There's an article on Healthcare IT Today with a lot of their details about why this happens and some ideas of how to get out of it, although it's a really challenging problem. But here's the thing that they said that was really interesting. 22:29.20 John Lynn sometimes the value doesn't actually happen until you scale it. And if that's the case, your pilot is not going to prove the ah roi the way you want. And so that makes it even more challenging for a healthcare leader that's looking at these solutions. 22:48.61 John Lynn But we have seen some, and that's the nice thing. We've seen people take and automate things that before would have taken forever. like Here's just one simple example that I saw with Ameritech Solutions. 22:58.41 Ratnadeep Bhattacharjee and 23:02.26 John Lynn They use UiPath, and they they automated the form filling for the state for this prenatal care that was needed to be provided. 23:04.72 Ratnadeep Bhattacharjee and 23:11.45 John Lynn The state requires, with certain situations, that all this prenatal care has to be documented. It's a 200-form field or something like that. right It's this outrageous form, Welcome to Government Work, that you have to fill out online for every... 23:25.73 John Lynn ah Every ah baby that's born, you know, to be able to show what's happening, and allows them obviously to do some public health ah around childbirth, which makes sense. They were spending hours and hours a day collecting the information and then putting it into the form. 23:44.04 John Lynn And now it's almost fully automated. the The form will take it from the EHR, all of the data. Now, it took a lot of workflow on their end to make sure that the er documented the data, that the people documented the right data in the EHR. 23:59.24 John Lynn And often, and it's more about in the right places in the EHR so that the agent knew where to get it because they might have documented it in a note, but this one should have been in the assessment or whatever, right? like It was in the wrong place, so the agent didn't know where it look. And so they had to refine that workflow. 24:15.00 John Lynn But now the agent knows where to get the data. It pulls the data. And then it goes, fills out the form for them and saves them so much time in being able to you know do this. So like what a stupid problem, right? like You're filling out a government form, but it saves them so much time. 24:31.01 Ratnadeep Bhattacharjee yeah 24:32.35 John Lynn So then that nurse can go and do something that's much more valuable than filling out a government form. 24:39.81 Ratnadeep Bhattacharjee is it this is This is a great example, John. One of the other things, John, that I, I mean, since this is this is the space that we operate on, right, know we operate it on the, you know, data integration, interoperability between those paradigms, right? 24:58.01 Ratnadeep Bhattacharjee So, John, where do you, especially for, you know, telehealth companies or virtual health companies, where do you see the ah evolution of EHR integration, claims data integration tools, ADD-fail integration tools, all these evolving, right? Because we used to have Merth Connect now over the last, you know very recently Merth Connect is also not open source as it was previously, right? And there are other platforms like Redox, Rhapsody, we have our own interoperability accelerator called SyncMesh. 25:32.05 Ratnadeep Bhattacharjee So how do you see that? market or not market, but how do you see the usability, real usability, real life scenarios of these solutions? Right. Do you see a shift or a change or how, how, how is it evolving? 25:47.91 John Lynn Yeah, well, it is interesting. Mirth, it was forked to Bridgelink, so there is an option there. And I know there's another one as well. So we'll see how that evolves. That's going to be an interesting ah you know open source versus private license ah thing. 26:03.92 John Lynn But you know what I think is even bigger than that is actually health care has this magical tool that no other industry has, which is the information blocking regulation. 26:16.91 John Lynn And we're just seeing some court cases around this that says you can't use it, you can't block the information, you need to make it accessible to me. you know There's a court case around data scraping that's along this lines, right? And is that okay? And is that the right way? And should, you know and I actually think that case, there's some other cases around HIE blocking and things like that, right? 26:40.18 John Lynn And when I look at those cases and I look at the information blocking rule, which I don't think they're going to back out on, turns out in the financial system, they had almost a similar rule that they already pulled away. 26:50.23 Ratnadeep Bhattacharjee Mm-hmm. 26:51.57 John Lynn But I think in health care, we're lucky. Amy Gleason's up there advising HHS and she's the acting Dodge administrator. She wants to data data sharing. In fact, I hear this week as we're recording it, 27:05.36 John Lynn They're supposed to have a meeting announcing all these vendors saying we're going to share data and we're going you know, I bet they're going to say we're going to get a full interoperable patient record, which is not a possible. We already talked about that. 27:17.05 John Lynn But anyway, they're going to say that, right? they're go to get They're committed to sharing. We'll see what comes of it, right? But so my point is, i don't think they're going to pull it back. I think information blocking is going to stay. 27:28.02 John Lynn And as you look at that, okay, if I'm an EHR vendor, watching these lawsuits, watching what's happening, i have to be thinking to myself, I better make this information available in a computable way through an API, through interoperability efforts, because if I don't, that could be problematic for me and my organization. 27:50.08 John Lynn So I'm hopeful that the largest EHR vendors There may be some ambulatory ones that don't quite want to go there or whatnot. It will be interesting to see how that plays out. But I think the largest ones are going to kind of see the writing on the wall and say, hey, if I don't open up my data, then I'm going to be forced to. 28:08.37 John Lynn And that could be even worse, right? 28:08.85 Ratnadeep Bhattacharjee Mm-hmm. 28:10.05 John Lynn So I'm hopeful that we see this evolution of EHR vendors saying, hey, I could be the system of record, but I don't have to be the application of record for every single thing that happens in my organization. 28:26.92 John Lynn So let's create the robust set of APIs that maybe even I could monetize to some degree, although I think that's part of the information blocking as well, how much you're allowed to. 28:38.12 John Lynn But at least you know I can pay for it, right, you know with people... you know paying me to use the APIs, and I can make this more accessible for entrepreneurs, for organizations who want to interact with the EHR, which is the system of record for healthcare, and to be frank, likely will be for quite a long time to come. 29:02.41 Ratnadeep Bhattacharjee Right, right. on ah On a separate note, ah John, so with the upcoming CMS mandates and all those regulations, especially on the Medicaid side of things, 29:14.19 Ratnadeep Bhattacharjee you know, they are decreasing the amount of money that's invested and whatnot, right? And also on the Medicare Advantage side of things, there there are a lot of, you know, regulations coming in and people are becoming really, you know, organizations, health plans are becoming really stringent and really strict in terms of making sure that all their risk models are perfect and, you know, how they kind of generate revenue via it becomes very important with all those regulations coming in, right? 29:41.76 Ratnadeep Bhattacharjee So how do you see the role of a golden record or a data warehouse or a data lake kind of an environment really shaping up in such an environment like a Medicare Advantage where where, you know, data is paramount in terms of making sure these models are perfect to the T, right? 30:03.05 John Lynn It's interesting, I've heard ah you know someone, in fact, a couple different people in the RCM space, the revenue cycle management space, they've they've said essentially this idea, which is payers have been investing in technology and data in order to understand how they're paying you. 30:23.70 John Lynn And then they say this next line, which is you as a provider need to do the same thing so that there's a level playing field when you're doing revenue cycle. And so I think that's the idea that's going to come forward, which is yeah, the payers are are becoming pretty advanced and in how they're approaching your data, understanding your data, understanding your trends and how they're going to pay you. 30:49.24 John Lynn and you need to do the same. So I think we see more and more organizations saying, i need the right data. the way i need to think like the payer. And if I'm going to think like the payer, then I do need the right data lake, the right data warehouse, the right you know data storage with the AI and other insights on top of it to help me understand what the payer is going to do. 31:16.19 John Lynn ah you know A simple example, right? You're about to file a claim. We see organizations starting to use the data to say, hey, when I file this claim, is it going to be denied? 31:27.98 John Lynn That way they're essentially acting like the payer, using technology like the payer would use to understand the claim and say, oh, yeah, it will be denied because it's missing XYZ. 31:38.95 John Lynn Right. And then you're like, oh, I can just go add XYZ. 31:39.41 Ratnadeep Bhattacharjee Hmm. Hmm. 31:41.96 John Lynn And then it runs through the filter. Oh, yeah, now it will be approved. Now you file it. So then you're you're essentially solving the claim denial before the denial even happens. And as they said, it's using the same technology that a payer would use, but using it on the provider side on the front end. So I think that's an example of what I think you're talking about, which is I need access to that data lake in order to know you know, how I'm going to be treated so that I can solve some of those problems early. 32:15.30 John Lynn And I think that's even going to become more important as we go into the value-based care area where people are going to be paid based on how they care for a patient population. 32:27.60 John Lynn Well, how do you care for a patient population? if you don't have data on that population, right? Like you can't, right? Now, to be fair, the data is only the first step. 32:37.72 Ratnadeep Bhattacharjee Thank you. 32:38.05 John Lynn The first step of the data is like, who should I interact with? What should I interact with them with? What are they at risk with, right? Like that's the data problem. There's still a secondary problem. Okay, now I know who the problem is. 32:49.84 John Lynn patients are, let's put it that way, where the risk is for me. But like, once I know the risk, I got actually solve the problem too. So it is going to be a multi-step approach to be able to do this effectively. 33:01.96 John Lynn But all of that starts with good data. Otherwise, I'm just, you know, hoping that, you know, some interaction is going to solve it. Like, i don't i don't think anyone wants to go to a value-based care contract with risk on both sides, even risk on one side without having the right data. to know how they're going to target and improve that population. 33:24.15 Ratnadeep Bhattacharjee No, completely, completely agree with you, John. Actually, the reason I asked you this is, you know, we have recently been engaged in a few projects and I can see a growing demand of good data warehouse, right? 33:38.28 Ratnadeep Bhattacharjee And wherein, you know, peers and these health systems, health plans rather already have some use cases in mind ah that can be on you know population health management side of things or you know risk side of things or whatever right but the paramount i mean the first thing is to set up a good data mart or a data warehouse whatever you want to call it right so i i see where you're coming from uh 34:06.02 John Lynn know what I don't think is going to happen though? I think we learned from before. I'm not just setting up a data warehouse to eventually solve some problems. I don't think we do that anymore. 34:18.09 John Lynn I think it's exactly what you said, which is we have these two problems we need to solve. What data do we need in the data warehouse to solve these two problems? And sure, if we can bring in some other data, that's great. We may import more than we need, right? 34:32.44 John Lynn Because we could potentially use it for something else. But I think we we learned, we start with the problem, we figure out what data is needed for that problem, and then we create the data warehouse to solve that problem, as opposed to a project of like, let's get all the data in the data warehouse, and somehow that's going to magically solve some problems we haven't defined. 34:44.96 Ratnadeep Bhattacharjee Right. 34:53.67 John Lynn So I think that's what's going to not happen. 34:56.90 Ratnadeep Bhattacharjee Yeah, I agree. That never works. You know, John, first thing we ask someone is what are you looking to build? 34:59.74 John Lynn Yeah. 35:03.37 Ratnadeep Bhattacharjee Data warehouse can be built. Everyone builds it. That's not the problem. 35:07.75 John Lynn Yeah. 35:07.75 Ratnadeep Bhattacharjee What do you want to build? What do you really want as an outcome? Because getting the data or the golden record as they call it in place is not the problem. The problem is what do you want to do with the data? Because that is essentially, ah that essentially will determine what kind of a data warehouse would you want to do? 35:24.11 Ratnadeep Bhattacharjee Because it's not a chicken and egg problem really. 35:24.53 John Lynn Yeah. exactly 35:26.31 Ratnadeep Bhattacharjee You know, people may call it that way. It's really what kind of outcome you want. 35:30.78 John Lynn ye 35:31.50 Ratnadeep Bhattacharjee So, so John, ah One of the other things I i mean on similar lines, right, on ah staying and the within the entire of inter interoperability ecosystem and paradigm. 35:44.41 Ratnadeep Bhattacharjee How do you see the evolution of FHIR or FHIR as they call it happening, right? FHIR, I mean, FHIR is touted to be something which can solve a lot of problems. 35:58.09 Ratnadeep Bhattacharjee ah and with all the regulations like TEFCA coming in and, you know, various mandates that are again coming, coming in and all the EHR vendors, some EHR vendors rather trying to embrace, you know, the concept of, you know, sharing data and, you know, making sure they're interoperable and whatnot. 36:17.69 Ratnadeep Bhattacharjee Where do you see FHIR really making an impact or is it really making an impact considering the kind of roadmap, you know, it had when it, was first you know introduced 36:29.91 John Lynn Well, I think it hasn't made the impact yet that it could and that I think it will. I think it's you know now finally mature enough that it's being rolled out in real applications. 36:44.42 John Lynn So we're starting to see the beginnings of what's going to be the impact of FHIR. And I think we're seeing it across very specific applications, you know smart on fire applications, things like that. 36:55.67 John Lynn But we're also seeing it in bulk fire, where they're taking bulk data using fire to solve certain problems. 37:00.03 Ratnadeep Bhattacharjee e 37:03.53 John Lynn And so I think both of those are mature enough now that we can actually leverage them. And so we're going to see that. And so I actually think you know we are seeing that shift to fire for certain situations that make sense. 37:20.70 John Lynn you know I say this with some minor trepidation because I read an article, i think it was a LinkedIn post you know by Brendan Keeler that highlighted something really interesting. and He says, every data standard makes the previous data standards look really stupid. 37:39.30 John Lynn Right. Like and fire did that to HL seven. Right. Like why would we want HL seven when we could do it on fire? 37:43.10 Ratnadeep Bhattacharjee Okay. 37:45.74 John Lynn Right. Like, because fire was so much more mature in the data quality and that it was the one offs. It was, you know, like you could query as opposed to, you know, just having it sent. 37:56.50 John Lynn Right. Like, you know, like all of that makes it interesting. And so. He's 100% right. Like, what will the next data standard be that will make FHIR look stupid? Like, why did we do all this FHIR structure that seems like so much work? 38:12.50 John Lynn If the next data standard, which I don't know if he, I can't remember if he hypothesized this or i did you know in my head, but I was like, yeah, the next standard could be an AI-driven pull the data and you already have all the data. You don't need the standardization. 38:25.94 John Lynn Because the AI already understands it and and does all the categorization for you. So it's like, ah you know, is that where we're headed? i don't know, right? I don't see it happening yet. But when you look at what AI can do, you're like, oh, yeah, why are we wasting all this time putting it into a standard format if AI could do it for us, right? So, ah you know, anyway, that so that's my, like, partial thing. He opened my eyes and I was like, oh, he's right. 38:52.26 John Lynn The only caution against that, though, is Every standard that we had, it took about a decade for the standard to become mature enough and adopted enough to actually make an impact. 39:05.22 John Lynn And so, you know, that was true with HL7, HL7v2, FHIR, you know, X12, you know, all of them, right? 39:05.47 Ratnadeep Bhattacharjee So 39:11.26 John Lynn Like they they all went through this... cycle and it's it's it's one part maturity and one part acceptance by the industry so that then the standard is incorporated into the software and other things that we do in order for it to be successful. So, you know, I look at that and I'm like, well, 39:30.56 John Lynn Is there another standard beyond fire right now? i don't think so. has been quicker, but does that mean we're still a decade decade away from another standard you know that does better than fire? 39:42.23 John Lynn Maybe, right? Maybe AI will make it five years. I don't know. But the point is, FHIR is the best we have for many situations. You could argue V2 is better for certain things, but X12 is, et cetera, right? like but But yeah, I don't see another standard on the horizon yet, and that's going to take years to implement. 40:02.56 John Lynn So I think if yeah we need FHIR to solve a lot of the problems that we have because it is the most mature standard that we have today and probably will be for the next five or more years. 40:13.60 Ratnadeep Bhattacharjee um Yeah, I remember the post by Brendan Kilmer by the way. ah so So, yeah, he's someone I follow as well. John, ah thank you. Thank you for sharing your unique perspectives on everything, right? AI, inter interoperability, you know, standards. 40:32.28 Ratnadeep Bhattacharjee You know, AI in healthcare cannot succeed without robust interoperability that is given. Your experience as both a media leader and advisor, you know, kind of gives our listeners a valuable look at the intersection of policy, innovation and real world adoption, right? 40:48.86 Ratnadeep Bhattacharjee Before we wrap up, John, where can our audience follow your work or connect with you online? 40:55.07 John Lynn Yeah, so I mean, the easiest way is go to healthcareittoday.com. That's our health IT community. We have a bunch of podcasts, videos, articles. You can subscribe to our email newsletter. you know I'm Tekka on Twitter, X, whatever you want to call it these days. I'm easy to find on LinkedIn. Just search for John Lynn. 41:12.76 John Lynn always love connecting with people so you can go there. 41:13.01 Ratnadeep Bhattacharjee Thank 41:15.58 John Lynn You know, I guess as the kids say, be sure to like and subscribe on ah all of our various channels. You know, it's always appreciative when people do that. And then if you're a healthcare marketing professional, we do have the Sway.Health community. 41:28.99 John Lynn That's Sway with two A's. So, you know, that's a great community of healthcare care marketers where we have similar things, a newsletter, videos, articles, but you know, like that. But for your audience, probably healthcare IT today is This the place to go to hear the latest in health IT. 41:44.51 Ratnadeep Bhattacharjee Right. Thank you. Thank you again, John. And to everyone listening, thank you for tuning in. If you enjoyed this episode, please subscribe and share it with your peers in healthcare and tech. This is Ratnati Bhattacharji and you have been listening to Leaders Perspective. 41:58.67 Ratnadeep Bhattacharjee Why healthcare AI needs interoperability to succeed. Until next time, let's stay curious and let's stay connected.
